For years, programmatic SEO sounded great in theory. Create pages at scale from repeatable patterns: same structure, different data. One page for every combination of product, service, or category. Indexed, ranked.
In practice, the result was almost always the same: template pages full of generic text that said nothing to anyone. Google caught on — and with the Helpful Content updates started penalizing exactly this type of content.
The idea wasn’t wrong. The execution was.
The real problem: scale without context
When a client — ESGRID — asked us to create hundreds of product pages for their site, we knew the traditional approach wouldn’t work. Many categories, many combinations, a massive number of pages to create. Doing it manually would have been unthinkable. But using generic tools meant producing content that knew nothing about the business it was being created for.
This is the core issue with programmatic SEO: scale alone isn’t enough. If what you’re scaling is generic, you’re just generating noise faster.
What was missing wasn’t the technology to produce content. It was context.
No tool knew who ESGRID was, what they sold, who they were talking to. And without that context, every output was interchangeable with any other business in the same industry.
The approach: a stack that knows the business
Instead of looking for one tool that did everything, we built a stack where each piece does one thing well and passes the result to the next. The ingredient that holds everything together is business context.
The custom tool (Lovable)
The part I’m most proud of is a tool I built in one afternoon with vibecoding, using Lovable.
It works like this: you feed it the client’s business profile — industry, business model, target audience, USP — and it generates specific keywords, blog article ideas, headlines and meta descriptions. Everything already calibrated to that business, not pulled from a generic database.
Four iterations from prompt to working product. It’s not a prototype: it’s the tool we use in production.
The tool connects to DataForSEO for real search volumes and CPC data. It uses Claude AI to generate article ideas calibrated to the industry. It also produces Google Ads copy consistent with the rest of the stack.
Cost per query: about $0.003. For hundreds of keywords, we’re talking a few dollars.
The automation (Make.com + Gemini)
The keywords generated by the tool flow into Make.com, which orchestrates the rest of the pipeline. Gemini receives the keywords along with the business context and generates the actual articles.
The key point: Gemini doesn’t just receive a keyword. It receives the keyword, the company profile, the tone of voice, the target audience. Every article that comes out is already contextualized.
The complete flow
Seed keyword
↓
Custom tool (Lovable)
→ Generates contextualized secondary keywords
→ Generates headlines, descriptions, article ideas
↓
DataForSEO API
→ Real volumes, CPC, difficulty
↓
Make.com + Gemini API
→ Generates complete articles with business context
↓
Custom tool
→ Generates consistent Google Ads copy
↓
Output: content + ads, calibrated to the businessWhat’s different from traditional programmatic SEO
The difference isn’t in the quantity of content produced. It’s in the relevance.
With generic tools, you generate 200 pages and then spend hours discarding, rewriting, adapting. 60-70% of the output is noise. With a stack that starts from business context, every output is already filtered. You don’t have to discard anything.
This has three practical consequences:
1. Actual production time is lower. It’s not just about how fast you generate content. It’s about how long you spend making it publishable. If the output is already relevant, total time drops dramatically.
2. Google treats the content differently. Pages that say something specific about a specific business perform better than templates filled with generic data. This isn’t theory — it’s what we see in GSC data after launch.
3. Cross-channel consistency. When the keywords, articles and Ads copy all come from the same context, the message is consistent. The client doesn’t need to manually realign anything.
When programmatic SEO makes sense
Not always. Programmatic SEO works when you have:
Repeatable patterns. Product/category combinations, service/city, feature/industry. If every page requires a completely different approach, it’s not programmatic SEO — it’s traditional content marketing.
Sufficient volume. If you need to create 10 pages, do them by hand. Programmatic SEO makes sense from 50+ pages up, where the stack setup cost gets amortized.
Structured data available. The tool works because it starts from concrete data — business profile, keywords with real volumes, product structure. Without this data, the output will be generic regardless.
Times have changed
Two years ago, building a stack like this would have required a dev team, a dedicated budget and months of work. Today all the pieces exist: Lovable to build custom tools without writing code, Make.com to orchestrate workflows, AI APIs for generation, DataForSEO for real data.
You don’t need to be a developer. You need to understand the problem, pick the right pieces and connect them.
Programmatic SEO isn’t dead. It’s becoming what it should have been from the start: content at scale that actually has something to say.
If you’re planning a content-at-scale project and want to know whether this approach is right for you, let’s talk.